Enhancing Meta Fuzzy Functions via Dynamic Pruning: A Nationwide Air Temperature Forecasting Framework for Türkiye

Abstract

This study proposes a novel Pruned Meta-Fuzzy Functions (PMFF) framework to forecast nationwide air temperatures across 81 cities in Türkiye. While existing literature often focuses on single models or standard hybridizations, these approaches frequently struggle with noise inherent in meteorological datasets and lack the necessary generalizability to perform across diverse topographical and climatic regions. To bridge this gap, the PMFF framework integrates a diverse pool of 10 base models that include statistical, machine learning, and deep learning architectures by systematically filtering out poorly performing and noise inducing predictors through a validation based Dynamic Pruning mechanism. Only the most reliable predictors are subsequently included in a Fuzzy C-Means (FCM) based meta-learning process to generate optimal forecasts. Experiments conducted on a comprehensive long-term dataset (2010–2026) demonstrate that the proposed PMFF architecture significantly outperforms traditional and state-of-the-art benchmarks. Rigorous statistical validation using the Friedman test and post-hoc Wilcoxon signed-rank tests with Holm-Bonferroni correction confirms that the performance gains are robust, statistically significant, and effectively mitigate forecast noise. Ultimately, this framework provides a resilient, highly accurate, and adaptive forecasting solution suitable for complex national meteorological applications.

Keywords

• air temperature forecasting
• meta-fuzzy functions
• dynamic pruning
• ensemble learning
• time series analysis

Meta-Bulanık Fonksiyonların Dinamik Budama ile Geliştirilmesi: Türkiye İçin Ulusal Hava Sıcaklığı Tahmin Çerçevesi

Abstract

Bu çalışma, Türkiye genelindeki 81 ilin hava sıcaklıklarını tahmin etmek için “Dinamik Budama” mekanizması ile güçlendirilmiş, özgün bir Budanmış Meta-Bulanık Fonksiyonlar (PMFF) çerçevesi önermektedir. Mevcut literatür genellikle tekil modellere veya standart hibrit yaklaşımlara odaklansa da bu yaklaşımlar sıklıkla meteorolojik veri setlerinde doğası gereği bulunan gürültü ile mücadele etmekte ve farklı topografik ve iklimsel bölgelerde performans göstermek için gereken genel geçerlilikten yoksun kalmaktadır. Bu boşluğu doldurmak amacıyla PMFF çerçevesi; istatistiksel, makine öğrenmesi ve derin öğrenme mimarilerini içeren 10 farklı temel model havuzunu, doğrulama tabanlı bir Dinamik Budama mekanizması aracılığıyla düşük performanslı ve gürültüye neden olan tahmincileri sistematik olarak eleyerek bütünleştirmektedir. Yalnızca en güvenilir tahminciler, daha sonra optimal tahminler üretmek üzere Bulanık C-Ortalamalar (FCM) tabanlı bir meta-öğrenme sürecine dahil edilmektedir. Kapsamlı ve uzun vadeli bir veri seti (2010–2026) üzerinde gerçekleştirilen deneyler, önerilen PMFF mimarisinin geleneksel ve güncel en iyi benchmark modellerini önemli ölçüde geride bıraktığını göstermektedir. Friedman testi ve Holm-Bonferroni düzeltmeli post-hoc Wilcoxon işaretli sıralar testleri kullanılarak yapılan titiz istatistiksel doğrulama, performans artışlarının sağlam ve istatistiksel olarak anlamlı olduğunu ve tahmin gürültüsünü etkili bir şekilde azalttığını doğrulamaktadır. Nihayetinde bu çerçeve, karmaşık ulusal meteorolojik uygulamalar için uygun, dayanıklı, yüksek doğruluklu ve uyarlanabilir bir tahmin çözümü sunmaktadır.

Keywords

• hava sıcaklığı tahmini
• meta-bulanık fonksiyonlar
• dinamik budama
• topluluk öğrenmesi
• zaman serisi analizi

References

1. [1] Koçak, H. (2023). Time series prediction of temperature using seasonal ARIMA and LSTM models. Gazi Journal of Engineering Sciences, 9(3), 617–629.
2. [2] Utku, A. (2024). Hybrid CNN-LSTM model for accurate long-term and short-term temperature prediction: A case study for Bingöl and Tunceli. International Journal of Pure and Applied Sciences, 10(2), 221–236.
3. [3] Utku, A., & Kaya, S. K. (2025). LSTM-based deep learning model for air temperature prediction. International Journal of Natural and Applied Sciences, 4(1), 55–68.
4. [4] Zhang, Z., & Dong, Y. (2020). Temperature forecasting via convolutional recurrent neural networks based on time-series data. Complexity, 2020, Article 3536572.
5. [5] Aghelpour, P., Mohammadi, B., & Biazar, S. M. (2019). Long-term monthly average temperature forecasting in some climate types of Iran using the models SARIMA, SVR, and SVR-FA. Theoretical and Applied Climatology, 135(1–2), 33–46.
6. [6] Curceac, S., Ternynck, C., Ouarda, T. B. M. J., Chebana, F., & Niang, S. D. (2019). Short-term air temperature forecasting using nonparametric functional data analysis and SARMA models. Environmental Modelling & Software, 111, 327–339.
7. [7] Elshewey, A. M., et al. (2023). A novel WD-SARIMAX model for temperature forecasting using daily Delhi climate dataset. Sustainability, 15(22), 15861.
8. [8] Karabulut, M. A., & Topçu, E. (2022). Derin öğrenme tekniği kullanılarak Kars ilinin hava sıcaklık tahmini. Mühendislik Bilimleri ve Tasarım Dergisi, 10(4), 1330–1342.

9. [9] Alomar, M. K., et al. (2022). Data-driven models for atmospheric air temperature forecasting at a continental climate region. PLoS ONE, 17(7), e0271517.
10. [10] Apaydın, M., Yumuş, M., Değirmenci, A., & Karal, Ö. (2022). Evaluation of air temperature with machine learning regression methods using Seoul City meteorological data. Pamukkale University Journal of Engineering Sciences, 28(3), 390–399.
11. [11] Kumaş, K., & Akyüz, A. Ö. (2024). Predicting Konya’s air temperature: Genetic programming, gradient boosting and random forest approaches. International Journal of Sustainable Engineering and Technology, 7(1), 36–49.
12. [12] Kumaş, K., Ayan, M., Akyüz, A. Ö., & Güngör, A. (2020). Antalya ili meteorolojik verileri yardımıyla hava sıcaklığının yapay sinir ağları metodu ile tahmini. Gümüşhane Üniversitesi Fen Bilimleri Enstitüsü Dergisi, 10(4), 1014–1024.
13. [13] Gong, B., Langguth, M., Ji, Y., Mozaffari, A., Stadtler, S., Mache, K., & Schultz, M. G. (2022). Temperature forecasting by deep learning methods. Geoscientific Model Development, 15(24), 8931–8956.
14. [14] Görücü, H. A., Özdemir, Y., Yıldırım, M., & Özdemir, G. (2026). Comparison of methods for weather temperature forecast in Vancouver City. Journal of the Institute of Science and Technology, 16(1), 1–15.
15. [15] Haque, E., Tabassum, S., & Hossain, E. (2021). A comparative analysis of deep neural networks for hourly temperature forecasting. IEEE Access, 9, 121754–121770.
16. [16] Wu, S., Shao, J., Li, G., Pu, X., Dong, Y., Du, T., Kang, J., & Zhao, B. (2026). TS-TCN: A deep learning model for ultra-short-term outdoor air temperature forecasting in complex terrain with small sample size. Energy and Buildings, 337, 116013.
17. [17] Tran, T. T. K., Bateni, S. M., Ki, S. J., & Vosoughifar, H. (2021). A review of neural networks for air temperature forecasting. Water, 13(12), 1654.
18. [18] Nketiah, E. A., Chenlong, L., Yingchuan, J., & Aram, S. A. (2023). Recurrent neural network modeling of multivariate time series and its application in temperature forecasting. PLoS ONE, 18(4), e0284795.
19. [19] Xue, X., He, X., Liu, B., Lyu, W., & Yang, M. (2025). Subseasonal forecasting and analysis of China temperature using an explainable deep learning model. International Journal of Climatology, 45(3), 1845–1862.
20. [20] Mutinda, J. K., Langat, A. K., & Mwalili, S. M. (2025). Forecasting temperature time series data using combined statistical and deep learning methods: A case study of Nairobi County daily temperature. International Journal of Mathematics and Mathematical Sciences, 2025, Article ID 9976314.
21. [21] Estrada-Patiño, E., Castilla-Valdez, G., Frausto-Solis, J., González-Barbosa, J. J., & Sánchez-Hernández, J. P. (2024). A novel approach for temperature forecasting in climate change using ensemble decomposition of time series. International Journal of Computational Intelligence Systems, 17(1), 1–19.
22. [22] Toharudin, T., Pontoh, R. S., Caraka, R. E., Zahroh, S., Lee, Y., & Chen, R. C. (2023). Employing long short-term memory and Facebook Prophet model in air temperature forecasting. Communications in Statistics - Simulation and Computation, 52(12), 6030–6047.
23. [23] Tak, N. (2018). Meta fuzzy functions: Application of recurrent type-1 fuzzy functions. Applied Soft Computing, 70, 342–356.
24. [24] Cifuentes, J., Marulanda, G., Bello, A., & Reneses, J. (2020). Air temperature forecasting using machine learning techniques: A review. Energies, 13(16), 4215.
25. [25] Grover, A., Kapoor, A., & Horvitz, E. (2015). A deep hybrid model for weather forecasting. In Proceedings of the 21th ACM SIGKDD international conference on knowledge discovery and data mining (pp. 379–386).
26. [26] Gong, Y., Zhang, Y., Wang, F., & Lee, C. H. (2024). Deep learning for weather forecasting: A CNN-LSTM hybrid model for predicting historical temperature data. arXiv preprint arXiv:2410.14963.
27. [27] Coutinho, E. R., Madeira, J. G., Borges, D. G., Springer, M. V., de Oliveira, E. M., & Coutinho, A. L. (2025). Multi-step forecasting of meteorological time series using CNN-LSTM with decomposition methods, Water Resources Management, 39(7), 3173–3198.
28. [28] Zaka, F., Nafisah, I.A., Lin, J., Almazah, M.M.A., Hussain, I., Almazroui, M., Chakrabortty, R., & Louati, H. (2025). Real-time temperature nowcasting using deep learning models across multiple locations, Modeling Earth Systems and Environment, 11(3), 165.
29. [29] Uluocak, I., & Bilgili, M. (2024). Daily air temperature forecasting using LSTM-CNN and GRU-CNN models, Acta Geophysica, 72(3), 2107–2126.
30. [30] Kalimi, K., Musyafa, A., Taryo, T., Sinambela, M., & Aji, T. W. (2026). Weighted ensemble of GRU, LSTM and XGBoost for multi-horizon temperature forecasting at a Tropical Highland Station, Journal of Applied Informatics and Computing, 10(4), 3576–3584.
31. [31] Mollick, T., Hashmi, G., & Sabuj, S. R. (2023). A perceptible stacking ensemble model for air temperature prediction in a tropical climate zone, Discover Environment, 1(1), 15.
32. [32] Oettli, P., Nonaka, M., Richter, I., Koshiba, H., Tokiya, Y., Hoshino, I., & Behera, S. K. (2022). Combining dynamical and statistical modeling to improve the prediction of surface air temperatures 2 months in advance: A hybrid approach, Frontiers in Climate, 4, 862707.
33. [33] Hyndman, R. J., & Athanasopoulos, G. (2021). Forecasting: Principles and practice (3rd ed.). OTexts.
34. [34] Winters, P. R. (1960). Forecasting sales by exponentially weighted moving averages. Management Science, 6(3), 324–342.
35. [35] Box, G. E. P., & Jenkins, G. M. (1970). Time series analysis: Forecasting and control. Holden-Day.
36. [36] Box, G. E. P., Jenkins, G. M., Reinsel, G. C., & Ljung, G. M. (2015). Time series analysis: Forecasting and control (5th ed.). Wiley.
37. [37] Chen, T., & Guestrin, C. (2016). XGBoost: A scalable tree boosting system. In B. Krishnapuram, M. Shah, A. J. Smola, C. Aggarwal, D. Shen, & R. Rastogi (Eds.), Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (pp. 785–794). Association for Computing Machinery.
38. [38] Breiman, L. (2001). Random forests. Machine Learning, 45(1), 5–32.
39. [39] Hoerl, A. E., & Kennard, R. W. (1970). Ridge regression: Biased estimation for nonorthogonal problems. Technometrics, 12(1), 55–67.
40. [40] Vapnik, V. N. (1995). The nature of statistical learning theory. Springer.
41. [41] Drucker, H., Burges, C. J. C., Kaufman, L., Smola, A. J., & Vapnik, V. N. (1997). Support vector regression machines. In M. C. Mozer, M. I. Jordan, & T. Petsche (Eds.), Advances in Neural Information Processing Systems (Vol. 9, pp. 155–161). MIT Press.
42. [42] LeCun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998). Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11), 2278–2324.
43. [43] Kiranyaz, S., Avci, O., Abdeljaber, O., Ince, T., Gabbouj, M., & Inman, D. J. (2021). 1D convolutional neural networks and applications: A survey. Mechanical Systems and Signal Processing, 151, 107398.
44. [44] Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural Computation, 9(8), 1735–1780.
45. [45] Bai, S., Kolter, J. Z., & Koltun, V. (2018). An empirical evaluation of generic convolutional and recurrent networks for sequence modeling. arXiv. https://arxiv.org/abs/1803.01271.
46. [46] Open-Meteo. (2026). Open-Meteo Weather API. Retrieved from https://open-meteo.com (Accessed June 1, 2026).
47. [47] Turkish State Meteorological Service. (2026). The state of Türkiye’s climate in 2025. Republic of Türkiye Ministry of Environment, Urbanization and Climate Change. Retrieved from http://www.emcc.mgm.gov.tr/files/State_of_the_Climate_in_Turkey_in_2025.pdf (Accessed June 1, 2026).
48. [48] Demšar, J. Statistical comparisons of classifiers over multiple data sets. J. Mach. Learn. Res. 7, 1–30 (2006